Shortly after the closure of the neural tube, a series of three vesicle-like swellings appear at the rostral apex of the tube. These swellings will eventually form the three primary brain vesicles: the prosencephalic (forebrain), the mesencephalic (midbrain), and the rhombencephalic (hindbrain) areas. Through various folding patterns the three primary vesicles ultimately give rise to the mature brain structures. The development of the three primary vesicles into their representative mature structures results in the following progression: the prosencephalic region develops into the telencephalon and the diencephalon; the mesencephalon remains undifferentiated, giving rise to the mature mesencephalon; and the rhombencephalon develops into the myelencephalon and metencephalon. Each of these secondary brain vesicles further develops into the recognised mature divisions of the human brain (Figs 2.5 and 2.6).

Figure 2.5 This figure demonstrates the development of primary and secondary brain vesicles.

Figure 2.6 Development of the secondary vesicles from the precursor primary vesicles. The prosencephalon differentiates into the telencephalon and the diencephalon. The mesencephalon remains undifferentiated as the mesencephalon. The rhombencephalon differentiates into the myelencephalon and the metencephalon.

The wall of the neural tube is formed initially by pseudostratified columnar epithelium that rests on the external basement membrane. Each cell sends a peripheral process to both the external (basal) lamina and the neural tube internal (luminal) surface. At the luminal surface a complex of junctional connections allows communication with other cells. The nuclei are found at different levels within the neuroepithelium, except those in the later stages of mitosis which are constantly found at the luminal aspect (Cowan 1981). The young neurons constantly migrate from the internal or luminal lamina to the external or basal lamina in a process called interkinetic nuclear migration. Throughout this process the neurons must continually retract and reform their peripheral processes. At this stage of development the neural tube consists of two histological areas of tissues, the ependymal or germinal cell layer found on the luminal aspect of the tube and the ventricular zone or matrix layer which spans the remainder of the tube and is in contact with the basal membrane. Continued cell division leads to growth and expansion of the neural tube in three ways: general expansion of the neural tube surface epithelium, rapid growth of the brain and spinal cord, and differentiation of the different lineages of cell types including glial cells and neurons. This differentiation phase is marked by a migration of some of the cells from the ventricular zone to the newly emerging intermediate or mantle zone, away from the luminal surface. The neurons in the intermediate zone send processes outward towards the basal membrane which eventually develops into the marginal zone. To summarise, at this point (4 weeks or 28 days) the development of the neural tube consists of four histologically identifiable layers: the ependymal or germinal layer, from which all the cells of the central nervous system will develop; the ventricular or matrix layer, which consists of newly formed cells undergoing interkinetic nuclear migration and mitosis; the intermediate or mantle zone, which consists of cells migrating from the intermediate zone; and the marginal zone, which consists of the elongated cytoplasmic processes of the cells in the intermediate zone. Interposed throughout these layers are developing glial cells including the radial glial cells which seem to act as guide cells for the migrating neurons (Fig. 2.7).

Figure 2.7 This figure illustrates the migration of neurons from the ventricular surface toward the plial surface of the neural tube. This migration takes place along highly specialised cells called neural glial cells. Note that neural migration appears to occur in cohorts, with neural cells born at or near the same time migrating at the same time.

Each population of neurons is generated during a distinct period of embryonic or fetal life. This period is usually quite short, ranging from a few days to a week. Cells born of the same cohort can be classified as embryological homologues. The relationships that each cohort of neurons forms with other neurons of the same cohort become very important as the system matures. These relationships can be used clinically to gain insight into the function of other related homologous systems by stimulating one system of neurons and measuring the response of a homologous system through a variety of clinical testing procedures (see Embryological Homologous Relationships later in this chapter). The sequence at which the cells withdraw from the proliferative pool is well defined. For example, in the retina the ganglion cells farthest from the original luminal zone are generated first, followed by the inner nuclear layer, which are in turned followed by the photoreceptors. The actual sequence may vary somewhat from region to region; for example, the situation in the motor cortex is completely reversed with the deepest cells forming first (Fig. 2.8).

Figure 2.8 A magnified diagrammatic view of Fig. 2.7. The various cell proliferation zones, the intermediate zone, and the migration of the neurons via the radial glial cells are illustrated.

The process of neurogenesis seems to be highly programmed in both space and time. The process seems to progress largely in a ventral to dorsal, and cervical to sacral gradient (above, down, inside, out). In keeping with this strategy, the motor systems arising from the basal plate located ventrally are generally generated before the sensory systems arising from the alar plate, which is located dorsally (Figs 2.9A and 2.9B).

Figure 2.9 A transverse section through the developing spinal cord of a human embryo 4 weeks old. (A) A schematic drawing of the real slide (B). Note the roof or alar plate and the floor or basal plate in (B).

It is now apparent that neurons and glial cells are generated simultaneously, and not in isolation as was once thought. Glial cells, however, tend to develop and proliferate long after neurons have been generated. This is evident in the formation of myelinated axons by the proliferation of glial cells subsequent to axon formation.

In a vast majority of cases the larger neurons of a system are generated before the smaller neurons of the same general type in the system. A notable exception to this occurs in the motor cortex where the smaller stellate cells of the cortex develop prior to the large pyramidal cells. The extent to which genetic determination plays a role in the fate of individual cells is unknown in mammals but thought likely to play a role.

Cells of the potential neuraxis that initiate their migration out of the ventricular or subventricular zones have achieved a rudimentary form of differentiation in that they are destined to become either neurons or glial cells. In the majority of cases, it is only after they have migrated to their desired destination and associated with other cells at their destination that they undergo their major transformation into neurons or glial cells.

Once started, the process of neuronal differentiation proceeds through three major classes of events, including morphological, physiological, and molecular differentiation.

How do the billions of neuronal connections that eventually form come to be? Are they formed randomly? Are they formed due to functional environmental input? Are they genetically predetermined? How do axons know where to go? These are the fundamental questions that investigators have been challenged with in neurobiology. As will be seen, the answers to these questions are very complex and probably involve a combination of the above possibilities at various phases of neuron development. Let us address each of these issues individually before considering a holistic view.

How do axons know where to go?

Developing ganglion cells in the inferior nasal portion of the retina send their axons to the lateral geniculate body of the thalamus, whereas developing ganglion cells of the superior temporal retina send their axons to the superior colliculus of the midbrain (Fig. 2.10).

Figure 2.10 The ganglion cell projections to the midbrain and thalamus from the retina. Note the anatomical relationship between the thalamus and the midbrain.

How do these neurons know where their respective axons are to go? In most developing embryos the paths taken by axons to their respective destinations is very constant and rigidly controlled. Even when axons are experimentally disorientated they still find their way to their target destination. This indicates that axons have some form of homing mechanism that allows them to know when they are going in the right direction and when they have arrived at the right location. Several mechanisms that allow axons to accurately find their way to their target destinations have been identified; these include selective axonal fasciculation; axon substrate interactions; axonal tropisms; and other gradient effects.

Newly formed axons from neighbouring cells often travel together over long distances using some form of axoaxonal connections to communicate. In many cases they follow a previously formed guide axon fibre with which they communicate in a similar axoaxonal fashion. This mechanism is termed axonal fasciculation.

Axons tend to grow in the direction that follows a selective substrate pathway specific for certain axonal membrane receptors contained on the surface of their developing axons. These molecules or receptors include various integrins such as fibronectin and laminin. Tropic influences include substances that promote axon growth along a concentration gradient. One such factor is nerve growth factor (NGF), which has been shown to exert strong growth influences on sympathetic nerve axons to the extent in some cases of causing them to change direction (Gundersen & Barrett 1979).

In light of the above discussion, it must be pointed out that even in fibre systems that seem to show high degrees of topographic order, such as visual systems, individual axons often diverge and follow pathways markedly different from those of neighbouring axons even when the destination is the same.

When the growth cone of an axon comes into close proximity of a postsynaptic cell surface at a potential target destination the terminal portion of the growth cone starts to accumulate vesicles. At the same time, morphologic changes occur on the pre- and postsynaptic membranes that allow the presynaptic transmitter to be recognised by the postsynaptic receptors. Functional synaptic integrity has been observed within minutes of the initial contact between an axon growth cone and a target muscle at acetylcholine (ACh) neuromuscular junctions (Kidokoro & Yeh 1982). Initially, the effect of the transmitter on the postsynaptic receptors is quite variable but as the functionality of the synaptic connection becomes stabilised, the action of ACh on the postsynaptic receptors results in a progressively shorter opening time of the sodium depolarising channels until a fairly consistent opening and closing time becomes established.

Initially, many more axons form synapses than are present in the mature system. Over time and through a variety of mechanisms, a portion of these axons are eliminated.

The mechanisms utilised to remove redundant or inappropriate axons are cell death and selective synaptic elimination.

Most neuronal systems undergo a phase of substantial neuron death at some phase of their development. In most neuron systems about 50% of the initial neurons formed undergo cell death. This process usually occurs temporally at the same time that the axons of the system have formulated contacts with their destination areas. This suggests that a certain amount of the stimulus for neuron death may actually arise or be initiated from the axon destination field through some form of feedback system (Hamburger & Oppenheim 1982). The feedback mechanism may be in the form of tropic growth factors produced at the destination site tissues. Active competition by axons for these growth factors may determine which axons and thus which neurons remain alive. In the case of dorsal root ganglion (DRG) cells one such growth factor that has been isolated is neuron growth factor, without which DRG cells cannot survive (Levi-Montalcini 1982).

Not only is there an overproduction of neurons initially but most neurons establish many more synaptic connections than are necessary or than they can physically maintain. This results in a phase of synaptic elimination in most systems. This was first recognised at the neuromuscular junctions where in the mature system each muscle fibre is innervated by a single axon. Early in development, however, many (6–7) axons may innervate a single muscle fibre (Perves & Lichtmann 1980). This same pattern has been shown to occur in many other systems including the autonomic system and is now thought to be a common strategy in most neuronal systems (Perves 1988). It is important to point out that axon collaterals may be eliminated but the parent axon and other synapses of the parent axon can remain functional and actually multiply in some cases. The most notable example of this occurs in the cortex. Initially, all pyramidal cells of lamina V in all cortical areas send axons through the cortical spinal tracts. In the case of the visual and other inappropriate areas of cortex, the inappropriate axons but not the parent cell bodies are eliminated so that only axons from the pyramidal cells of the motor and some areas of sensory cortex remain in the cortical spinal tracts at maturity (O’Leary & Stanfield 1989).

Whether a particular synapse or axon collateral remains and is not eliminated seems to depend on the degree of stimulation generated at the postsynaptic membrane. Synapses that generate a response frequently in the postsynaptic membrane develop a stronger connection with the postsynaptic region which ensures their continued existence. This relationship between stimulation and circuit stability is a form of neural plasticity.

Spina bifida occulta is a defect of the posterior arch of a vertabra in which one or the other of the developing pedicle segments fails to fuse to form the spinous process. In spina bifida occulta failure of the arch formation does not affect the development of the thecal sac or its contents. The most common areas of the spine involved are the lumbosacral areas. Clinical manifestations of spina bifida occulta usually only become apparent sometime after birth and include back pain, increased incidence of disc herniation, and spondylolisthesis (Fidas et al. 1987; Avrahami et al. 1994). Although neurological manifestations are rare, a number of conditions have been associated with spina bifida occulta. These include low termination of the conus medullaris (tethered cord syndrome), syrinx formation, lipoma, nerve root adhesions, and conjoined nerves (Giles 1991; Gregerson 1997).

In situations involving spina bifida vera there is a wide bony defect in the posterior arch development of usually more than one vertebra. The thecal sac and its contents are usually also involved and protrude beyond the confines of the spinal canal. There is some evidence to suggest that an adequate supply of folic acid during this critical period of development can prevent this type of condition. Failure of fusion of the posterior arch to the degree necessary to result in spina bifida vera must take place in the period of the 21st to 29th fetal day. Unfortunately, this is a period in which most women do not realise they are pregnant; thus to be effective supplementation with folic acid must begin prior to conception. Herniation of the fluid-filled sac that contains cerebral spinal fluid is called meningocele. When protrusion includes the meninges, cerebral spinal fluid, and neural elements, it is called a myelomeningocele. When neural elements protrude without thecal covering, it is called a myelocele (Figs 2.14A and 2.14B). Myeloschisis refers to the presence of complete uncovering of the neural elements along a sagittal midline defect that involves bone, thecal sac, and all posterior tissues (Guebert et al. 2005) (Fig. 2.15). Failure of closure of the caudal neuropore results in absence of the cranial vault with the cerebral hemispheres either completely missing or reduced to non-functional masses. This condition is referred to as anencephaly (Figs 2.16A and 2.16B).

Figure 2.14 (A) The different variations of spina bifida vera. There is a wide bony defect in the posterior arch development of usually more than one vertebra. The thecal sac and its contents are usually also involved and protrude beyond the confines of the spinal canal. (B) The appearance of spina bifida vera in an infant.

Figure 2.15 Myeloschisis.

Figure 2.16 Failure of the caudal neuropore to close. (A, B) This condition results in absence of the cranial vault with the cerebral hemispheres either completely missing or reduced to non-functional masses. This condition is referred to as anencephaly.